ADL: autodock Digest, Vol 182, Issue 12

Van Dam, Hubertus hvandam at bnl.gov
Wed Apr 15 11:55:12 PDT 2020


Hi Venkatesh,

I concur with Diogo, we have successfully used Autodock with GNU parallel on our HPC clusters over the last few weeks for exactly the reasons you mentioned. The nice thing with GNU parallel is that it automatically takes care of load balancing the work with a node. Starting tasks in the background and using a "wait" statement, as an alternative, does not automatically adapt to different tasks taking a different amount of time.

Best wishes,

    Huub

-----------------------------------------------------------------------------------------------------
Hubertus van Dam, 631-344-6020, hvandam at bnl.gov
Brookhaven National Laboratory
 

On 4/15/20, 13:51, "autodock-bounces at scripps.edu on behalf of Diogo Martins" <autodock-bounces at scripps.edu on behalf of diogo.stmart at gmail.com> wrote:

    Hello,
    
    Please note that HDC, OAA, OAC OAHD and OACC are not standard atom types.
    Unless you created these atom types, there's something wrong.
    
    I recommend GNU parallel. If you have a text file with all the docking
    commands, one command in each line:
    
    $ cat commands.list
    autodock4 -p dock1.dpf -l dock1.dlg
    autodock4 -p dock2.dpf -l dock2.dlg
    ...
    autodock4 -p dockN.dpf -l dockN.dlg
    
    GNU parallel can execute these single-threaded processes in parallel:
    $ parallel -j 8 < commands.list
    
    Best regards,
    
    On Wed, 15 Apr 2020 at 04:00, venkat <azbvenkat at gmail.com> wrote:
    
    > Hello,
    > Thank you very much Martins  and Hubertus ,  Now I am able to perform the
    > autogrid and autodock. As per your guidance, I rerun the script
    > prepare_dpf  with option -p ligand_types, for that I used following
    > command.
    >  *pythonsh  ../Utilities24/prepare_gpf4.py -d TRI_PLPRO_COM_LIN/ -r
    > plpro.pdbqt -p npts="40,40,40" -p gridcenter="-19.198,53.911,-34.169" -p
    > ligand_types="C Cl HD HDC OA OAA OAC OAHD OACC NA SA" -o grid.gpf*
    >
    > After grid generation I modified affinity map according to the
    > ligand_types.
    >
    > I doubt/clarification, How to perform Autodock in HPC clusters (let say
    > utilize all processor which available in machine), because, through script
    > I submitted the job it only tacking one processor out of 8. Following
    > script I used to submit job
    > #!/bin/csh
    > cd /Dockings
    > ls  >2docking.list
    > foreach d (`cat 8docking.list`)
    > echo $d
    > cd $d
    > autodock4 -p $d.dpf -l $d.dlg
    > cp $d.dlg ../../dlgs/
    > cd ..
    > end
    >
    > kindly  provide your valuable answer
    > Thank you
    >
    > Sincerely
    > venkatesh
    > ________________________________________________
    > --- ADL: AutoDock List  --- https://urldefense.com/v3/__http://autodock.scripps.edu/mailing_list__;!!P4SdNyxKAPE!Q_wVkd7PRxyIczrFutavpoTEbgO86CAlMIZEAxAe6UVN4Mvxn8vM6V3Ku5HJ5uI$  ---
    >
    >
    ________________________________________________
    --- ADL: AutoDock List  --- https://urldefense.com/v3/__http://autodock.scripps.edu/mailing_list__;!!P4SdNyxKAPE!Q_wVkd7PRxyIczrFutavpoTEbgO86CAlMIZEAxAe6UVN4Mvxn8vM6V3Ku5HJ5uI$  ---
    
    




More information about the autodock mailing list